The proposed project is designed to clarify, expand upon, and apply growth mixture modeling as it relates to understanding developmental pathways that lead to adolescent drug abuse and dependence. Growth mixture modeling techniques are recently emerging quantitative methods that permit investigators to identify discrete developmental patterns of change over time. These new methods hold great promise for the social sciences, arid especially for the study of pathways of drug use, because they enable investigators to construct taxonomies of normal and maladaptive developmental patterns. Each developmental pathway of risk may have a unique etiology or special portent for later psychopathology. Because little is known about these emergent methods, the primary goal of this proposal is to review and evaluate growth mixture models for their relevance to research in developmental psychopathology and drug abuse. The project is organized around three specific aims. Aim 1 is to review and contrast two recently developed growth mixture modeling techniques relative to one another and to more traditional analytic models. Aim 2 is to examine the behavior of growth mixture models with data simulated to reflect variations and conditions that would commonly be encountered in applied research on drug use and abuse. The results from Aims 1 and 2 will be applied in Aim 3, which provides a pedagogical demonstration of growth mixture modeling with real world data by testing the specific hypothesis that particular developmental pathways of antisocial behavior will predict adolescent drug abuse.